Hi,
I just noticed, that Spider allows, when providing actuals, to have Activity duration to be independent of resource work duration.
In the example below the activity has finished earlier, than resource has stopped working on it.
Is is something which has been missed in the software or this is done on purpose?
I seem not to be able to think of real life situation, which would be modeled by such behavior of the software.
In the example below it also creates a problem, because Activity 2 has 2 days lag after Activity 1, which is not being obeyed in this case (assuming Actuvty 1 duration in reality shall be equal to duration of resource work on this activity)
https://docs.google.com/file/d/0B1FBt_G3gCVqTW5MRzVQNElyeVE/edit
Evgeny,
we discuss actual data. And in my example with meeting I don't know what time will take the travel for all meeting participants (and I did not mean meeting with the customers).
Spider Project is used for management of all consruction projects (about 300) for preparation of 2014 Winter Olympic Games. We helped with the management of this portfolio. In one of reports it was written that some bridge is 90% ready and the photo was attached with cars moving on this bridge in both ways. When I asked why this bridge is only 90% ready when it is functioning, the answer was that everything is ready except getting the formal permission to build this bridge.
It is not right but it happens. We do not plan this way but permit people to enter actual data the way it really happens. It could not be calculated. If activity is actually finished and following activity can be performed, remaining work on finished activity may be entered as resource assignment duration.
Vladimir,
I agree, that people shall be able to report the actual duration and actual start and finish dates of activities. My only comment is that if people have already started working on Activity, doesn't it simply mean, that activity started and if they work, after activity has finished, doesn't it mean, that activity has not finished yet in reality?
This is interesting concept, which you put in Spider, but I just wonder, how it works on practice. In case of Spider somebody needs to take an arbitrary decision, to mark the task as finished, whilst the people are still working on it.
I agree, that if activity is finished to the limit, that it's successors can start, most probably it can be marked finished (this is what also the document from Rafael suggests), however a new activity shall be created for remainder. This activity (e.g. Finish Remaining Items of the Bridge) should have it's own predecessors, successors, resources etc., in another words it shall be treated as any other activity (this is at least how I do it in my schedules).
I think this is also better for historical records. In most of the cases, if somebody started working before activity start or continued working after it's finish, this means, that there was a problem in initial planning. In this case it is better either to extend activity or create a separate one with explicit explanation of what this new activity is all about. In this case if somebody will analyze the historical records with the goal to create a schedule for similar project, he will notice, that in reality it took 5 review-updates cycles to finish document, not just 1st time right.
By allowing some work on activity after it has finished, I think you make such historical analysis difficult as it hides the actual picture.
In case of travel to meeting, the initial "Meeting with the customer" event is better to have initial duration, which takes into account travel. Otherwise Spider may schedule somebody to be in the office in Chicago next day after one day meeting in Moscow, whilst in reality it will take one day to travel back.
Regards.
Evgeny
P.S. I know you probably have plenty of improvement requests from real customers, but my be it could be a good idea to implement a switchable option "Allow actual activity duration to be independentl actual of assignments durations"
Vladimir,
What you say it makes sense to me, just look at the following reference.
http://www.nflaace.org/index_files/john_orr_cost_loaded_schedule_updating_pdf.pdf
Spider allows for The Great Divorce.
I believe we shall look for ways to keep the historic records in practical ways that in case of a claim after an analysis can tell us what really happened but without unneeded granularity nobody is to follow.
Entering on the CPM database resource hours at the individual resource level can be too much, it is enough to report man-hours per activity, but man-hours for every named resource is too much. Also it might be better to report actual hours per update period but not the per-hour detail, leave it to the payroll records. Even if using a time card software it shall condense the data to manageable scale at the activity level.
On a Condominium Building we usually have a cement plaster activity per floor, say 20 floors 20 activities, but we want to track cost for a single cost account, same as hours, if for 20 individual activities is too much, just imagine tracking hours per named resource on each activity, even per trade is too much. At times your plan is for 10 masons, 10 mason helpers, 4 laborers but in actual execution the composition of the crew varies every day. Keeping such level of detail on a CPM just because the database can handle it is nuts. Weekly total man-hours for labor resources per activity shall be enough while for future work planning we still need the detailed crew composition.
We usually update our schedules at end of the month but at the cutoff date we do not have resource hours until the following week payroll period ends, the same goes for other costs, many entered on the system weeks after, when invoice is received. Most of us do not implement job costing at the PO but at the Invoice, very frequently they do not match and making adjusting entries on the system would be too much. We should be able to update schedule for status and time projections purposes but latter to add some costs and resource hours after schedule update. The data entry screen for each purpose shall be simplified, a single table with hundreds of columns drives me crazy.
Always keep it simple enough, otherwise no one will implement it. Needless to say this posting represents a change of mind from my prior postings, to me a search for constant improvement that will never end.
Best Regards,
Rafael
Evgeny,
we do not want to ban an opportunity to report that somebody started to work on activity before its start and worked even after it was finished. This can be inputed only manually and will be seen.
Imagine that an activity is a meeting and somebody added time for travel to and from the meeting. What is wrong with this? A person was really busy before and afer activity dates and I don't think that it makes sense to add special activities for each person travel.
By default Spider does not do it but it permits to enter such data manually. This is about actual data, so just check it.
Best Regards,
Vladimir
Vladimir,
thanks for explanation:
RE:
<<
There is certain sequence for entering actual information: from the left to the right and from the assignments to activity
>>
So, when it comes to duration of activities, the following behavior is observed in Spider (it also follows your logic of sequence for entering actual information, mentioned above):
But I think common sense says, that it shall be opposite:
But I can see, that at the moment Spider does not reinforce this logic
Or am I missing something?
Regards.
Evgeny
Yes, there is a freedom for our users to enter actual information that differs from planned.
By default activity and resource duration are the same but in fact resources may spend more or less hours than planned and work on the task after the result was submitted or finish the work but submit the resullts later. Entering actual resource information is advanced function. Usually people enter activity actual data and resource data are automatically adjusted. But if real work was done with deviations people may enter actual data that differs from planned.
There is certain sequence for entering actual information: from the left to the right and from the assignments to activity. But again,Spider does not recalculate actual data using planned calendar. People may report that they worked through the night in case of emergemcy and enter more hours that exist in planned calendars. This shall be controlled by their managers.
how do you do this on your real projects? Do you just manually check, that this misalignment between resource work and project duration is never entered by those responsible?
Re "The fact that resources and activities can have different calendars and that actual performance might have broken scheduled work time makes it to me a bit complicated"
I agree with you, that calendars shall be taken into account only when we are talking about the activities ahead of Data Date (in the future). When it comes to actually performed work, tool shall allow to overwrites this. Just because the reality always overwrites plans.
Rafael
Rafael,
how do you do this on your real projects?
Do you just manually check, that this misalignment between resource work and project duration is never entered by those responsible?
Re "The fact that resources and activities can have different calendars and that actual performance might have broken scheduled work time makes it to me a bit complicated"
I agree with you, that calendars shall be taken into account only when we are talking about the activities ahead of Data Date (in the future). When it comes to actually performed work, tool shall allow to overwrites this. Just because the reality always overwrites plans.
Regards.
Evgeny
Evgeny,
I agree with you, I am trying to find a logic on how it shall be done. It shall be easy to program the software to automatically change the activity dates to prevent resource usage outside activity range.
I am having issues on how to date activity duration shall be calculated and how this impact remaining duration calculations. The fact that resources and activities can have different calendars, that actual resource performance might have broken scheduled work time and that activities might progress without resource loading makes it to me a bit complicated.
For a start I would like to understand actual algorithm for such calculations.
Best Regards,
Rafael
Rafael,
I am not sure I understood from your answers whether you consider the behavior, described by me, to be a problem or there is a good reason for it.
Specifically, I don't understand, what is the reason to be able to have activity duration shorter, than the duration of the actual work of a resource, which is being done under this activity
Regards.
Evgeny
Evgeny,
Not only you can edit resource finish to finish after activity finish but also to start before activity start.
An update record at the Actual Data Input Table consists of one activity line to report progress under no resource presence and a separate line for every resource usage being reported.
I believe that maybe:
Hope this helps to communicate some ideas fresh from the oven, not fully cooked, one more turn and back to the oven.
Best Regards,
Rafael